Farnaz Jahanbakhsh (@[email protected])
@FarnazJ_
Followers
1K
Following
2K
Media
10
Statuses
258
Assistant prof at @UMichCSE, PhD @MIT_CSAIL, Postdoc at @StanfordHAI | HCI, Social Computing
Joined January 2021
My work was featured in the MIT News today!.Read more on empowering users to assess the accuracy of content and see assessments from their trusted sources everywhere on the web:
1
1
39
📢 I'm recruiting PhD students for fall 2024 to join my new research lab at @UMichCSE!.Interested in social computing systems that empower people? Consider applying!. Please spread the word 📢.
19
134
430
📢📢I'm recruiting PhD students for Fall 2025 to join my research lab at @UMichCSE!.Interested in doing research in social computing systems? Consider applying! 📢📢.
1
99
325
2 (rather overdue but still very exciting) announcements: I’ll be joining @UMichCSE as an assistant professor in fall 2024! And this year, I’ll be a postdoc at @StanfordHAI working with @sanmikoyejo and @msbernst!. Beyond thrilled for these next steps 😊.
33
9
267
PhD Defense ✅.Incredibly lucky and grateful to have been advised by @karger and to have had @arvindsatya1 and @AdamBerinsky on my dissertation committee!
32
1
138
I’m on the job market! I design and build systems that reimagine the design of social information spaces (SM platforms and the web) to increase end-user empowerment and reliability of online content. Please get in touch if you know of any relevant industry or academic positions!.
4
24
101
📢 Excited to share our upcoming #CHI2023 paper. The result of an internship @IBMResearch and collaborating with my awesome mentors @dakuowang , @michael_muller, @lucian_popa_us, and Yannis Katsis.1/n
3
6
83
Struggling to fall back asleep when I gain the slightest amount of consciousness, lying in bed debating for hours if I should check the news on Iran, knowing that I won’t be able to sleep if I do—or if I don’t.
4
3
64
Thanks for making my time here so wonderful! I'll miss you all so much! ❤️.
Our good friend @FarnazJ_ is going off to @UMichCSE as an Assistant Professor!. We are so lucky to have been able to share time with Farnaz-- she is a brilliant and talented researcher, a fun and empathetic friend, and overall great person to be around ❤️ We will miss her dearly!
4
1
53
By a show of hands, who is going to #CHI2023? 🙋♀️ Looking forward to meeting new friends and catching up with old ones!.
2
1
47
Journalist friends, we have built 2 browser extensions that we hope can help you in your quest for delivering truthful content to people. We’d like your help in evaluating them through a user study: Please r/t. 1/3.
1
19
28
Caught between wanting to take a break from all this distress and yearning to be aware of what is happening; wanting to sleep as a refuge and fearing that I would wake up to a pile of atrocities that I’d endure better if I learned about them in bits.
0
0
22
There was a thread a while ago with HCI researchers discussing and being very open about how much their consulting hourly rates are. I can't seem to find it though. Does anybody have a pointer to it? Or do you mind sharing yours?.
1
0
19
I was given a Schengen visa for Germany and my flight has a layover in the Netherlands. When checking in at @Delta, they told me because I have a type D visa, I can transit through the Netherlands but I can't find such a rule anywhere online.
2
0
16
@UMichCSE You can take a look at some of the work I've done here The topics will not be limited to these, but they can give you an idea of the type of research we’ll do. Remember to apply by December 15!
0
1
16
Less than one week left to apply!.
📢📢I'm recruiting PhD students for Fall 2025 to join my research lab at @UMichCSE!.Interested in doing research in social computing systems? Consider applying! 📢📢.
0
4
14
Our #CSCW2022 paper ⭐️Understanding Questions that Arise When Working with Business Documents⭐️ will be presented tomorrow (Nov 9, 14:30 ET) and Nov 18. The result of an amazing internship with @adamfourney @ryen_white @ellinoox & Robert Sim @MSFTResearch.
0
2
13
Got excited and tried Aragorn, Gondor, Anor, and Grond one after another until I had to accept that a day may come when the #NYTSpellingBee accepts LoTR proper nouns, but it is not this day.
0
0
10
Do you ever come across content (esp images or videos) online that you find unpleasant or even anxiety-inducing? Fill out this survey to help us understand how to help you and get a chance at a raffle for $30 gift cards: Please r/t.
0
9
11
We're asking crowd workers to do a longish task where they differentiate between items (say articles). We're worried that fatigue might inhibit differentiation. We want to introduce a break where they do sth lightweight and fun. Has anyone tried this? What activities did you use?.
3
1
10
It used to be that I’d check the news to relax and take my mind off work. Now I work to relax and take my mind off the news.
1
0
9
There's still one more day to submit your paper!.
1st Workshop on Reliable Evaluation of LLMs for Factual Information (REAL-Info):
1
0
8
Very grateful for the time I got to spend here and the chance to work with and get to know you all! ❤️.
@FarnazJ_ STAIR will miss you too! Amazing to have you here, and best wishes for your next steps.
0
0
9
Argh those #BigBangTheory scenes when Sheldon or occasionally Leonard get annoyed that someone has ended their sentence with a preposition annoy me to no end. And then I have to pause the show and annoy everybody else with an account of English grammar.
0
0
7
I recently posted a survey on certain subreddits to capture the perspectives of a certain user population that I could only reach on Reddit. Participants would be entered in a raffle for gift cards. When going through the responses, I identified *89%* of the entries as spams. 1/n.
3
0
6
Uh. What?.
A bill is filed in Texas legislature to ban citizens, governments & entities of China, Iran, North Korea & Russia from purchasing land in Texas. I will sign it. This follows a law I signed banning those countries from threatening our infrastructure.
0
0
4
not just on social platforms. It is unrealistic to expect that every website on the web, some the very perpetrators of misinformation, offer misinformation moderation and in a fair and rigorous way. And here’s where our CHI'24 paper comes to the rescue.
0
0
5
@drmaxlwilson @EricPSB @_JimWallace @AghajariZhila One issue is that the response is supposed to include not only the changes made, but also arguments for why certain comments or concerns by a reviewer do not apply to the work and therefore have not been incorporated in the revision and that's hard to do in such a limited space.
1
0
4
@syardi CATME ( is a tool for both assigning students to teams based on instructor-specified criteria as well as peer evaluation in class team projects.
0
0
5
Had this dream where I was fighting an alien with a shovel and after they got wounded I repented and fetched herbs to heal them. But because when you reconcile with an alien you morph with them into one being, for me to exist independently they had to fork me into a child process.
0
0
5
While reading Frederic and Elfrida from Austen’s Juvenilia, I had to flip back to the front cover 3 times, gobsmacked, to make sure this was really written by #JaneAusten.
0
0
4
Oh and I forgot to mention that after establishing that the type is written elsewhere, the rep pointed at the C written for type and exclaimed “see, here’s the D” 🤦♀️.
0
0
4
For years I refrained from rating any movie that wasn’t LOTR. But Don’t Look Up has so many undeserved and motivated negative reviews that I just had to.
0
0
4
So after spending an entire hour at @Delta's counter and/or with customer service online, I asked where they on the visa they were looking to see the type, at which point they pointed at the country abbreviation D next to Deutschland.
1
0
4
What’s the purpose of this AI you ask? It could serve as an accuracy aide to the user, scoping their attention to posts that need their attention. Or it could serve as a guard—when the user is about to share a post and is thinking more about likes than accuracy. 5/n.
1
0
3
It gives any user the power to assess content accuracy & specify whose assessment they trust. Of course platforms are unlikely to change their design so drastically, not w/o push from the outside. What’s more, people encounter (mis)information anywhere on the web,.
1
0
4
@hen_drik This paper tests the efficacy of a number of strategies ranging from attention checks to more sophisticated approaches 👉 Though I've seen free-text survey responses that look suspiciously like they were generated by LLMs, and that's a different beast.
0
0
3
Ok I discovered Wordle like 5 minutes ago and I can already see how my addiction to it will unfold. 202 4/6.🟩⬜⬜⬜⬜.🟩⬜⬜🟨⬜.🟩🟩🟩⬜🟩.🟩🟩🟩🟩🟩.
2
0
4
I eventually could check in just in time for boarding once we established that my visa was type C. Although I still can't find what the problem with transiting with a type D visa would have been.
1
0
4
@karger Check out our upcoming CSCW paper exploring democratized assessments and trust relationships to combat misinfo on social media: and our other CSCW paper about empowering users to modify and "fix" news headlines:
0
2
3
But wait, can displaying a user’s personalized AI’s prediction of how the user is likely to assess content create a self-fulfilling prophecy? I’m thinking this is a good cliffhanger to leave you on. So how about you read the paper to find out 🙂.6/6.
0
0
2
Has anyone encountered similar issues when administrating surveys? How do you deal with them? 4/4.
3
0
3
@HelenasResearch @msbernst @RanjayKrishna @tobigerstenberg @jennwvaughan @QVeraLiao @bansalg_ @adamfourney @arvindsatya1 Congrats, that's amazing!! ❤️🥳.
1
0
3
And here this CHI paper comes in. We explore a personalized AI that is trained on a user’s assessments of content and predicts how the user would assess other content that they haven’t assessed. The AI evolves and updates its predictions as the user assesses more content. 4/n.
1
0
2
@IBMResearch @dakuowang @michael_muller @lucian_popa_us In my research, I argue that social platforms need to enable users to assess content accuracy, similar to how they capture likes and comments. And that users should have a say in whose assessments they want to see. 2/n.
1
0
2
Now a problem is that there’s just SO much content— and so many posts that are similar. It’s a shame that a user’s assessment of a post should be tied to the post and not shown on similar content. 3/n.
1
0
1
One lets any user assess any page on the web as accurate or inaccurate, or ask about its accuracy. Whoever follows the user and opens the page or comes across a link to the page will see their assessments right on the page. 2/3.
1
0
2
Some of the submissions actually looked legit. But the giveaway was that they were 1. either *verbatim* the same as multiple other submissions, or 2. rephrasing of other responses with conspicuously similar words in their response to every question except for demographics. 2/n.
1
0
2
0
0
0
0
0
2
@yuwen_lu_ @karger @amyxzh @LeaVerou @mrdoob @MIT_CSAIL Yes! The Reheadline extension repo (.interfaces with this backend ( which is a larger system that manages a trust network for accuracy assessments of content in addition to reheadlines.
1
0
2
Here's the workshop link: For some reason the link in the original tweet is not working.
0
0
1
Oops, the next week's presentation will be on Nov 15th. Nov 16th was misinformation 😅.
0
0
2
My daily word game ritual is getting out of hand: #Nytspellingbee, #Wordle, and now the better Wordle (/ht @emhastings317):.
And a first version of Lordle of the Rings is available: /ht @gwjschenk for suggesting 'Lordle'./ht @youyuxi for creating vue-wordle (and vue.js!).and of course /ht @powerlanguish for Wordle in the first place!.
1
0
2
0
0
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
1
0
0
1
@shagunjhaver Ah that's a really neat idea! But yeah it probably wouldn't work for projects that aim to understand the experiences of certain groups of people with platforms other than the one where they're being recruited.
0
0
1
0
0
1
0
0
1
Ok told ya. 203 3/6. ⬜⬜🟨🟨⬜.🟩🟩🟩🟩⬜.🟩🟩🟩🟩🟩.
0
0
1
@Shaheer17663563 @UMichCSE Yes, that'd help both me and the student figure out if we're a good match.
0
0
1
@LeaVerou Isn't the idea behind invalidating the session after a while reducing the probability of session hijacking?.
1
0
1
0
0
1
0
0
1
0
0
1
The other extension identifies news headlines on any article’s page and lets you submit alternative headlines for the article. Whenever those who follow you come across the headline anywhere on the web, they will see your suggested alternative headline. 3/3.
0
0
1
0
0
1
0
0
1
The survey start times for these datapoints are within a few seconds of each other. The IP addresses are different. I figured if the demographics info for these submissions is not valid, the same is true of the other responses so I discarded them all. 3/n.
1
0
1
0
0
1